Cisco Agent Builder: Securing AI Agents from Supply Chain Attacks

Security researchers recently uncovered a sophisticated malware operation called SmartLoader. Attackers cloned a legitimate server connecting AI assistants to Oura Ring health data. They then submitted a trojanized version to legitimate registries. This was a live attack, successfully poisoning registries used by real developers. Developers searching for a health data integration could have installed credential-stealing malware without warning. This highlights a growing supply chain risk with the deployment of new AI agents. Cisco is addressing this by building AI Defense directly into Cisco Agent Builder. This makes it the first enterprise AI agent platform with native security at every lifecycle stage. Security is now part of the platform itself, from scanning third-party integrations to inspecting agent execution in real time. Cisco Agent Builder is part of Cisco Cloud Control Studio. It allows enterprises to turn third-party tools, operational knowledge, and workflows into AI agents. These agents can then act across their entire environment. This development matters because it aims to protect businesses from increasingly complex AI-related cyber threats.